Which governments moved from AI plans to practical capacity?
BCG's 2026 comparison places Japan and South Korea in its leading group and says eight economies improved their readiness tier. The index maps capacity and exposure; it does not show that AI policy has improved public outcomes.
The short version
- What happened
- Japan and South Korea join Canada, mainland China, Singapore, the United Kingdom and the United States in BCG's highest readiness tier.
- Who it affects
- Workers and public-service professionals are more likely to feel the difference between strong and weak readiness through training, infrastructure, procurement and appeal processes than through a country's rank alone.
- What remains unproven
- This is a consultancy-authored composite index, not a causal evaluation of government policy.
Keep in mind
This is a consultancy-authored composite index, not a causal evaluation of government policy.
The 33 indicators use different years and sources, and the report provides no confidence intervals for close scores.
Public-company earnings calls and LinkedIn job data may under-represent informal work and less digitally visible economies.
Source analysis · 1 linked source. Primary analysed source date: 20 October 2026.
Read the full report · 4 min The evidence, methods, implications and limitations
The leading group widened—and several smaller economies moved up
Japan and South Korea have joined Canada, mainland China, Singapore, the United Kingdom and the United States in the top group of Boston Consulting Group's 2026 AI Maturity Matrix. BCG also reports that Chile, the Czech Republic, Ecuador, Greece, Japan, Lithuania, Nigeria and South Korea moved to a higher readiness tier. The result matters because it suggests that practical capacity is not confined to the largest technology economies, although the matrix ranks inputs and conditions rather than lived outcomes.
The report covers more than 75 economies and uses 33 indicators. It first estimates how exposed each economy is to AI across 26 sectors, then compares that exposure with readiness. Higher exposure is not automatically better or worse: it can indicate room for productivity gains, but also greater disruption risk if skills, safeguards and institutions lag. BCG therefore groups economies by whether their readiness appears ahead of, aligned with or behind their exposure.[1]
Readiness is a composite of six different capabilities
The readiness side uses BCG's ASPIRE framework: ambition, skills, policy and regulation, investment, research and innovation, and ecosystem. Exposure combines BCG estimates of task-automation potential, Quid analysis of earnings calls from more than 5,000 public companies and LinkedIn data on AI-related job postings as of 20 April 2026. The three exposure measures are normalised, combined and weighted by each sector's contribution to national output.
Most input data are from 2024 to 2026, although some series are older; the startup measure spans 2013 to 2024. That makes the matrix a broad comparative snapshot, not a live dashboard. Equal weighting choices, gaps in cross-country data and the use of public-company and LinkedIn signals can favour economies with more visible formal-sector activity. The report provides rankings and tiers but no confidence intervals showing when small differences are statistically meaningful.[1]
The distinction that matters
How the matrix turns national signals into a readiness comparison
The index compares economic exposure with six kinds of enabling capacity; it does not directly score social benefit.
Exposure
Combine sector, company and job signals
Estimate where AI may matter most
Readiness
Score six ASPIRE dimensions
Map capacity to respond
Tier
Compare readiness with exposure
Flag alignment or gaps
Governments can use the gaps as a checklist, not a league-table verdict
For public officials, the useful question is which capability constrains a country's response. A government with strong research but weak digital infrastructure may need deployment capacity; one with investment but limited training may need workforce support; one with high exposure and thin rules may need oversight. For workers and public-service users, those distinctions affect whether new systems arrive with training, appeal routes and reliable infrastructure rather than as isolated pilots.
The report does not test whether a higher tier produces better productivity, safer systems, improved services or more widely shared gains. It is also authored by a consultancy that advises governments and companies on AI, and the public report does not publish a full sensitivity analysis for alternative indicator weights. The next evidence should connect specific policy changes with transparent outcomes—service quality, firm adoption, wage and job effects, incident rates and distribution across regions and groups. For now, the defensible finding is that measured capacity is broadening, while implementation quality remains unproven.[1]